Simulation of Ocular Drug Distribution Following Administration of Intracameral Bimatoprost SR Using 3D Computational Fluid Dynamic Model

Abstract

Purpose : To predict human ocular tissue drug concentrations of bimatoprost (bim) following intracameral administration of a biodegradable sustained-release (SR) bim implant (BimSR) using 3-dimensional (3D) computational fluid dynamic (CFD) modelling.

Methods : A 3D dog ocular CFD model that accounts for convection and diffusion of drug in various ocular compartments, including uptake by the iris-ciliary body (ICB), a purported site of action of bim, was developed. Drug diffusion coefficients in aqueous humor and vitreous were estimated using bim molecular weight and published methods. Drug uptake by choroidal vasculature was modelled as a sink term in the drug transportation equation. Drug concentration was set to zero in the trabecular meshwork due to convectional drug uptake driven by aqueous humor flow. The in vitro drug release profile from the implant was used to estimate release rate and flux. Bim concentrations were simulated in the dog model over 110 days and compared with experimentally determined aqueous humor, cornea, and ICB concentrations following BimSR 15 μg administration to beagle dogs. Simulations were extended from dog to human using an existing 3D human ocular CFD model and by setting the aqueous humor flow rate using published data to simulate human ocular bim concentrations after BimSR administration.

Results : Bim concentrations predicted by the 3D dog ocular CFD model were in agreement with experimental data for aqueous humor (all timepoints), and the cornea and ICB at days 52 and 80. The 3D human ocular CFD model was corroborated by clinical data: the predicted human aqueous humor bim concentration was in agreement with that measured in BimSR-treated patients in a Phase 3 trial. In the human model, simulated bim ICB concentration was below the FP receptor EC₅₀ for IOP lowering effect after 106 days post-BimSR administration whereas the median IOP lowering effect in humans was 9 months after 1 implant.

Conclusions : Simulation results in dog and human were corroborated by experimental and clinical data. Human model results suggest the bim concentration in target tissue falls below the EC₅₀ after 106 days post-BimSR administration. Continuous targeted drug delivery yields sustained IOP lowering beyond the duration of drug presence, likely due to durable tissue remodeling of aqueous outflow pathways.

This is a 2020 ARVO Annual Meeting abstract.